Table 4.
Multiple logistic regression models of factors predictive of mortality and morbidity
| Factor | Dependent variable: mortality | Dependent variable: morbidity | ||||||
|---|---|---|---|---|---|---|---|---|
| aOR | p | 95% CI | aOR | p | 95% CI | |||
| Severity of shock | ||||||||
| MAP < 60 (or non-palpable BP) | 8.42 | <0.001 | 3.13 | 22.66 | 4.83 | 0.002 | 1.80 | 12.94 |
| MAP 60 or higher | 1 | 1 | ||||||
| Parity | ||||||||
| 5 or more live births | 1.33 | 0.35 | 0.73 | 2.42 | 2.43 | 0.04 | 1.06 | 5.58 |
| 0-4 live births | 1 | 1 | ||||||
| Primary Diagnosis | ||||||||
| Uterine atony | 1.44 | 0.19 | 0.83 | 2.49 | 2.68 | 0.07 | 0.93 | 7.76 |
| Other condition | 1 | 1 | ||||||
| Where bleeding began* | ||||||||
| Transferred in bleeding | -- | -- | -- | -- | 1.82 | 0.51 | 0.30 | 10.93 |
| Began bleeding at RH | -- | -- | -- | -- | 1 | |||
| Study Phase | ||||||||
| NASG | 0.45 | 0.004 | 0.27 | 0.77 | 0.20 | 0.002 | 0.07 | 0.56 |
| Pre-intervention | 1 | 1 | ||||||
NASG = Non-pneumatic Anti-Shock Garment. aOR = adjusted odds ratio. Reference groups for categorical variables shown in italics. Hospital facility included as control variable in both models, but not shown in Table 4. The number of observations in Table 4 is less than 1442 because of missing data; n = 1038 for the morbidity model, and n = 1442 for the mortality model. Robust standard errors used to adjust for clustering at the facility level.
* Where bleeding began was not a significant predictor of mortality, but it was associated with morbidity, in bivariate analysis. Therefore it is included in the multiple logistic regression model of factors predictive of morbidity only.